2021 IEEE International Conference on Acoustics, Speech and Signal Processing

6-11 June 2021 • Toronto, Ontario, Canada

Extracting Knowledge from Information

2021 IEEE International Conference on Acoustics, Speech and Signal Processing

6-11 June 2021 • Toronto, Ontario, Canada

Extracting Knowledge from Information

Technical Program

SPE-32: Speech Recognition 12: Self-supervised, Semi-supervised, Unsupervised Training

Session Type: Poster
Time: Thursday, 10 June, 13:00 - 13:45
Location: Gather.Town
Session Chair: Ebru Arisoy-Saraclar, MEF University
 
SPE-32.1: HUBERT: HOW MUCH CAN A BAD TEACHER BENEFIT ASR PRE-TRAINING?
         Wei-Ning Hsu; Facebook AI Research
         Yao-Hung Hubert Tsai; Carnegie Mellon University
         Benjamin Bolte; Facebook AI Research
         Ruslan Salakhutdinov; Carnegie Mellon University
         Abdelrahman Mohamed; Facebook AI Research
 
SPE-32.2: A FURTHER STUDY OF UNSUPERVISED PRETRAINING FOR TRANSFORMER BASED SPEECH RECOGNITION
         Dongwei Jiang; Didi Chuxing
         Wubo Li; Didi Chuxing
         Ruixiong Zhang; Didi Chuxing
         Miao Cao; Didi Chuxing
         Ne Luo; Didi Chuxing
         Yang Han; Didi Chuxing
         Wei Zou; Didi Chuxing
         Kun Han; Didi Chuxing
         Xiangang Li; Didi Chuxing
 
SPE-32.3: PRE-TRAINING TRANSFORMER DECODER FOR END-TO-END ASR MODEL WITH UNPAIRED TEXT DATA
         Changfeng Gao; Key Laboratory of Speech Acoustics and Content Understanding
         Gaofeng Cheng; Key Laboratory of Speech Acoustics and Content Understanding
         Runyan Yang; Key Laboratory of Speech Acoustics and Content Understanding
         Han Zhu; Key Laboratory of Speech Acoustics and Content Understanding
         Pengyuan Zhang; Key Laboratory of Speech Acoustics and Content Understanding
         Yonghong Yan; Key Laboratory of Speech Acoustics and Content Understanding
 
SPE-32.4: SEMI-SUPERVISED SPEECH RECOGNITION VIA GRAPH-BASED TEMPORAL CLASSIFICATION
         Niko Moritz; Mitsubishi Electric Research Laboratories (MERL)
         Takaaki Hori; Mitsubishi Electric Research Laboratories (MERL)
         Jonathan Le Roux; Mitsubishi Electric Research Laboratories (MERL)
 
SPE-32.5: UNSUPERVISED DOMAIN ADAPTATION FOR SPEECH RECOGNITION VIA UNCERTAINTY DRIVEN SELF-TRAINING
         Sameer Khurana; Massachusetts Institute of Technology
         Niko Moritz; Mitsubishi Electric Research Laboratories (MERL)
         Takaaki Hori; Mitsubishi Electric Research Laboratories (MERL)
         Jonathan Le Roux; Mitsubishi Electric Research Laboratories (MERL)
 
SPE-32.6: IMPROVING STREAMING AUTOMATIC SPEECH RECOGNITION WITH NON-STREAMING MODEL DISTILLATION ON UNSUPERVISED DATA
         Thibault Doutre; Google Inc.
         Wei Han; Google Inc.
         Min Ma; Google Inc.
         Zhiyun Lu; Google Inc.
         Chung-Cheng Chiu; Google Inc.
         Ruoming Pang; Google Inc.
         Arun Narayanan; Google Inc.
         Ananya Misra; Google Inc.
         Yu Zhang; Google Inc.
         Liangliang Cao; Google Inc.